Google DeepMind, Google's flagship AI research lab, on Monday announced Veo 2, a next-generation video-generating AI and the successor to Veo.

Veo 2 can create two-minute-plus clips in resolutions up to 4k (4096 x 2160 pixels), which is 4x the resolution and over 6x the duration OpenAI's Sora can achieve, according to the company.

But for now, in Google's experimental video creation tool, VideoFX, where Veo 2 is exclusively available, videos are capped at 720p and eight seconds in length.

Google said it's expanding the number of users accessing VideoFX this week.

"Over the coming months, we'll continue to iterate based on feedback from users," said Eli Collins, vice president of product at DeepMind.

Like Veo, Veo 2 can generate videos given a text prompt or text and a reference image. It can also more realistically model motion, fluid dynamics, and properties of light. That includes different lenses and cinematic effects, as well as "nuanced" human expression, according to DeepMind.

"Coherence and consistency are areas for growth," said Collins. "There's also room to improve in generating intricate details, fast and complex motions, and continuing to push the boundaries of realism."

To mitigate the risk of deepfakes, with the proprietary watermarking technology, SynthID, it embeds invisible markers into frames Veo 2 generates, Deepmind said.